<< 06-06-2019 >>

---

00:05:15	*	envoyt quit (Ping timeout: 248 seconds)
00:05:37	*	envoyt joined #nim
00:19:44	*	Jjp137 quit (Ping timeout: 248 seconds)
00:22:12	*	Jjp137 joined #nim
00:29:04	*	stefantalpalaru quit (Quit: stefantalpalaru)
00:35:42	*	deech quit (Ping timeout: 272 seconds)
00:46:03	*	deech joined #nim
00:51:13	*	oprypin_ is now known as oprypin
01:12:27	*	sealmove joined #nim
01:20:19	*	deech_ joined #nim
01:29:34	*	c3v0axz quit (Ping timeout: 246 seconds)
01:30:18	*	c3v0axz joined #nim
01:36:14	skrylar[m]	alright, its up on github for snooping https://github.com/Skrylar/thunderstrike
01:39:13	*	Snircle quit (Quit: Textual IRC Client: www.textualapp.com)
01:42:41	*	deech quit (Ping timeout: 258 seconds)
01:45:31	*	envoyt quit (Ping timeout: 248 seconds)
01:55:32	*	Snircle joined #nim
02:02:30	*	Snircle quit (Remote host closed the connection)
02:11:39	*	theelous3_ quit (Ping timeout: 248 seconds)
02:23:08	*	envoyt joined #nim
02:30:23	*	dddddd quit (Remote host closed the connection)
02:32:39	*	Amun_Ra joined #nim
02:35:19	*	deech_ quit (Ping timeout: 252 seconds)
02:35:35	*	laaron- quit (Remote host closed the connection)
02:37:31	*	leorize quit (Ping timeout: 256 seconds)
02:37:53	*	laaron joined #nim
02:38:49	*	leorize joined #nim
02:47:44	shashlick	@Araq @narimiram - nightlies passed on devel, need to backport the fixes into version-0-20. https://travis-ci.org/genotrance/nightlies
03:02:03	*	tiorock joined #nim
03:02:03	*	tiorock quit (Changing host)
03:02:03	*	tiorock joined #nim
03:02:03	*	rockcavera quit (Killed (verne.freenode.net (Nickname regained by services)))
03:02:03	*	tiorock is now known as rockcavera
03:11:41	*	chemist69 quit (Ping timeout: 258 seconds)
03:13:37	*	chemist69 joined #nim
03:21:35	leorize	shashlick: with 0.20 csources, you can now build for all os/arch with make and enjoy parallelization :D
03:22:51	shashlick	should we update CIs?
03:23:54	leorize	I've covered everything except appveyor
03:24:30	leorize	that should bring down csources build time to <30s (from exp. on azure)
03:24:58	*	I_Right_I quit (Remote host closed the connection)
03:25:06	leorize	can you update appveyor? :) %NUMBER_OF_PROCESSORS% should be used for make's -j parameter
03:26:51	shashlick	does the makefile make it into the xz file as well?
03:27:34	leorize	build.sh does that?
03:28:11	leorize	oh, then idk
03:30:05	shashlick	ideally the same makefile also builds nim - just builds koch if it is there and then koch boot and tools
03:30:32	shashlick	the all a person would have to do is make and it would make csources and nim for the average downloader
03:50:37	sealmove	guys I am making a hex editor, it's almost ready, I am so hyped :)
03:55:46	sealmove	sneak-peek https://asciinema.org/a/guTVkURUY3Cl95fNllflDczzR
03:56:58	leorize	kitty user :)
03:57:48	sealmove	how do you know!?
03:58:12	sealmove	ah, asciinema has it as metadata
03:58:58	leorize	I like kitty, but it doesn't support iosevka so I dropped it :p
03:58:59	sealmove	well, I only use kitty because it renders images nicely, so it's easy to enable that feature in ranger
03:59:35	leorize	build a rival to alacritty in Nim :D
03:59:58	*	nsf joined #nim
04:00:20	sealmove	f*ck yeah!
04:04:35	sealmove	But first is the hex editor. I didn't find any nice one, except r2, but oh well, you have to navigate in it to reach the hex editing mode and even then the defaults are weird (for example the offset might not be at 0).
04:05:24	sealmove	Then there is hexyl, but it's only a hex dumper, not editor.
04:40:27	*	brainproxy quit (Ping timeout: 248 seconds)
05:06:49	rayman22201	I briefly tried to contribute to alacritty... Cool project, but it was one of the reasons I gave up Rust for Nim lmao
05:13:10	sealmove	What was the reason?
05:13:33	rayman22201	The Crate system
05:13:59	rayman22201	The crate system + cross platform development
05:14:14	rayman22201	I was trying to get windows support to work
05:19:18	sealmove	If I understand correctly, the problem is that Rust's standard library is too small and crates vary in quality/stability?
05:20:28	*	leorize quit (Quit: WeeChat 2.3)
05:24:39	*	narimiran joined #nim
05:27:10	rayman22201	No. (Though that was more true back when I used it. There are many more crates these days.) The problem I had was that configuring crates to deal with cross-platform use was very complicated and error prone.
05:27:35	rayman22201	And there were many implicit defaults that I found odd
05:38:01	sealmove	So basically you are saying the problem is that crates are inconsistent with one another?
05:39:04	sealmove	While languages like C# and Nim have more consistency because there are guys (guy? :P) making calls?
05:39:18	rayman22201	yes, but I did not like the configuration system itself for crates.
05:39:54	rayman22201	Crates are very consistent and very opinionated. Too much so. If you have to make anything that they haven't accounted for, then you are screwed.
05:40:10	rayman22201	Like say, which windows kernel version you using...
05:41:09	rayman22201	I don't need a whole yaml file + pragmas, I need typesafe ifdef...
05:41:59	rayman22201	Note: this was ~2 years ago. Things may have changed / gotten better since then.
05:42:31	sealmove	I see, so basically hard to hack/adjust crates
05:42:36	rayman22201	Also, having "sub-crates" was a PITA.
05:42:52	rayman22201	dependencies in the same repo basically.
05:43:47	rayman22201	The story isn't much better in Nim land in this regard, but I also don't feel forced into making "sub-nimble" packages either, so it doesn't bother me as much.
05:44:40	rayman22201	Rust relies soooo heavily on Crates, that you are forced into it, even if you don't want it.
05:45:03	sealmove	Well it forces you to do tons of things
05:45:17	sealmove	To work in a very particular way
05:45:20	rayman22201	I think it over-complicated the architecture of alacritty from an early state, and it was unnecessary complexity.
05:46:06	rayman22201	But even hello world forced me to make a Crate (which had even more boilerplate back then). I find that absurd...
05:46:37	rayman22201	Yeah. Nim let me just "get shit done"...
05:46:58	sealmove	Yes I feel the same way. It feels overarchitectured, though coming from Mozilla I can see their reasons. For really large projects it having such a rigid workflow is probably good.
05:47:47	rayman22201	Yeah, I understand the reasoning. But It should be something that a project grows into, not is forced into from the beginning.
05:49:46	sealmove	Consistency vs simplicity
05:50:07	sealmove	Rust doesn't trust programmers at all :>
05:50:18	rayman22201	Exactly. lol
05:50:49	rayman22201	It's not like I don't "feel safe" when using Nim. lol
05:51:11	rayman22201	But I also don't feel like the compiler treats me like an idiot.
05:51:42	rayman22201	Rust just always thinks you are a terrible programmer. I am terrible sometimes, but as much as Rust thinks lol.
05:52:07	sealmove	Yeah I think that's the gist of it
05:52:33	sealmove	Products like Rust and Go are aimed for noobs
05:52:52	sealmove	And they just get in the way of a normal person, lol
05:53:18	sealmove	I understand big companies like Mozilla and Go want to feel protected from stupidity
05:53:47	sealmove	But come on... tools shouldn't replace knowledge/skill
05:54:08	rayman22201	Well, from a business standpoint, it makes sense. It's cheaper to hire crappy programmers and give them safety scissors.
05:54:32	Araq	I doubt it.
05:54:46	Araq	crappy programmers always do more harm than good
05:55:11	rayman22201	Lol. I agree. Java already tried this strategy, and look at all the terrible Java code in the world these days :-P
05:55:16	Araq	but yeah, you cannot hire only PhDs to do the programming
05:55:22	rayman22201	too bad the business people didn't seem to get the message.
05:55:32	rayman22201	Since they keep trying.
05:56:12	rayman22201	Also true. We can't yet clone Araq or make AraqML :-P
05:56:26	sealmove	Well, at the very least this makes Rust a very unfun language imo. I don't get why people get hyped with it, maybe it's marketing.
05:56:37	*	leorize joined #nim
05:56:57	rayman22201	It's all marketing. It's one of the best marketing I've ever seen for a programming language.
05:58:37	Araq	nah, I'm sorry, Rust is to be respected.
05:59:01	Araq	They solved many many problems.
05:59:31	Araq	And even if you disagree with their answers, it asked the right question
06:00:15	rayman22201	ok ok. I'll admit that I'm over-exaggerating. I agree with you on that point Araq.
06:00:25	rayman22201	They did ask the right questions.
06:01:01	rayman22201	And they opened the door to allow other newer languages to flourish.
06:04:01	Araq	now ... where is my =move implementation?
06:04:13	Araq	clybber?
06:08:50	*	leorize quit (Remote host closed the connection)
06:15:38	skrylar[m]	rust is still that way
06:16:11	skrylar[m]	or i should say cargo is still that way
06:17:18	*	leorize joined #nim
06:33:16	*	PMunch joined #nim
07:00:00	*	gmpreussner quit (Quit: kthxbye)
07:04:30	*	gmpreussner joined #nim
07:04:36	*	laaron quit (Remote host closed the connection)
07:05:41	*	krux02 joined #nim
07:06:29	*	Amun_Ra quit (Quit: brb)
07:07:45	*	laaron joined #nim
07:10:28	*	Amun_Ra joined #nim
07:14:25	*	Vladar joined #nim
07:24:29	FromGitter	<xmonader> how is rust for noobs? :D
07:25:03	FromGitter	<xmonader> I agree on Go being simple, too simple language, but rust?
07:29:18	*	leorize quit (Quit: WeeChat 2.3)
07:37:10	*	laaron quit (Remote host closed the connection)
07:38:42	*	laaron joined #nim
07:56:19	*	floppydh joined #nim
08:08:47	*	Tongir joined #nim
08:40:40	FromDiscord_	<sealmove> xmonader: Because it doesn't trust you to do things your way.
08:46:09	FromGitter	<xmonader> It's more of helping you avoiding common problems, that kept happening over the years no?
08:46:40	FromDiscord_	<sealmove> If you are experienced, you trust yourself and don't want the language to make decisions for you. That's why C++ people won't pick it.
08:47:31	FromDiscord_	<sealmove> It mostly inexperienced programmer and gets in the way of experienced ones.
08:47:40	FromDiscord_	<sealmove> mostly helps*
08:47:55	FromGitter	<arnetheduck> rust and well-written c++ code are nearly identical.. the experienced programmers probably don't notice the difference ;)
08:48:16	FromGitter	<xmonader> C++ is dumpster for all features, and yet is isn't it offering ways to mimic move semantics now? and bunch of 1000 types of pointers
08:49:09	FromGitter	<arnetheduck> well, the experienced c++ will avoid the old features, and indeed use moves and all that stuff to never allocate memory, to have the compiler deal with ownership and resource mgmt etc
08:51:20	FromGitter	<arnetheduck> likewise with threading, et al.. sure, it's hard to reach that level of proficiency because of all the baggage, but once you manage to slim it down, it'
08:52:06	FromDiscord_	<sealmove> The real burden is... Trying to do low level things yourself. It's nearly impossible to implement a nontrivial data structure in Rust, so you always get it from a crate. The actual implementation is a unsafe Rust {}
08:53:39	FromGitter	<arnetheduck> well yeah, that escape hatch is needed - compilers are not omniscient and haven't beaten turing on undecidability yet so..
08:55:02	FromDiscord_	<sealmove> So I think it defeats the purpose. Because if you can write good unsafe Rust then you don't need protection while writing simpler code. That's why I said it's for less experienced people.
08:55:25	FromGitter	<arnetheduck> the selling point is that those parts of your code are easy to find simply becuase of how they stand out - the search space for those classes of bugs is dramatically reduced.. I'd say that's the driving reason for its popularity - we see that a lot in spaces like crypto and networking where security demands are high
08:57:31	FromGitter	<arnetheduck> nope, you can be as good as you like, but I'll trust the compiler (or really, the theory behind it) more than your attempt to be faithful to the principles that you need to follow.. humans are not that great at the kind of attention to detail, but they are great at overestimating their own capabilities... "can't be wrong" beats "shouldn't be wrong in my opinion"
08:57:39	FromDiscord_	<sealmove> Yes, if you want 100% guarantees, it's a good choice, but it's too niche.
08:58:23	FromDiscord_	<sealmove> The added complexity is not worth it most cases.
09:00:03	FromGitter	<arnetheduck> would actually be nice to have some nice first-class support for bridging nim and rust, to gain access to all that code that adheres to the stricter guidelines that the language requires, and where the developer has been forced to thing up-front about a whole bunch of stuff
09:01:47	ehmry	if rust, why not ada/spark then? I've never seen a successful rust compile, but I know that ada works
09:04:32	ehmry	it seems like people choose rust because its a fad, not because of assurance, we've had better high-assurance stuff than rust for while now
09:08:34	Araq	fad is one aspect, the other is that Ada is much better for embedded devices with less memory and worse for when you can afford a heap
09:09:17	federico3	ehmry: care to expand on "we've had better high-assurance stuff than rust"?
09:09:21	ehmry	ada
09:09:39	FromGitter	<xmonader> because rust checks other boxes too, like being bit modernish and some fancy fp styles. I'd pick a lisp without gc, that what inspired rust after all no?
09:09:43	ehmry	ada had a spec before a compiler, rust has compilers but no spec
09:10:22	Araq	and Ada has like the worst marketing of all time, for instance, Ada has 2 stacks, this means you can return variable-length data *from* a function without touching the heap, it's gold for the "allocations are slow" crowd. unfortunately they don't know it
09:14:24	FromGitter	<xmonader> June is here and V lang is going to render all languages obsolete :D
09:17:07	livcd	well the site is down
09:17:45	*	laaron quit (Remote host closed the connection)
09:18:03	FromGitter	<xmonader> @livcd nim's?
09:19:28	*	laaron joined #nim
09:23:43	livcd	vlang
09:26:55	Zevv	https://github.com/vlang/v/blob/master/examples/tetris/tetris.v first time I see some actual code
09:27:44	*	PMunch quit (Remote host closed the connection)
09:30:05	FromGitter	<xmonader> definitely better syntax than go, yet no enums :(
09:30:07	*	Tongir quit (Ping timeout: 246 seconds)
09:33:09	*	abm joined #nim
09:33:13	*	stefanos82 joined #nim
09:34:06	*	PMunch joined #nim
09:36:51	FromDiscord_	<sealmove> What about Rust as a compile target?
09:37:15	FromDiscord_	<sealmove> Nim could compile to Rust to ensure guarantees.
09:37:43	FromGitter	<xmonader> or Nim can get the move semantics :D
09:38:31	FromDiscord_	<sealmove> Just saying, maybe Rust is an interesting target for compiling to.
09:38:48	Araq	the only guarantee that Rust offers and Nim doesn't is the guarded pointers. which can't Rust infer all the time, so guess what, Nim couldn't infer it either
09:39:19	Araq	and so we can either compiler to unsafe Rust or we need lifetime annotations on our own
09:39:29	Araq	which kind of defeats the purpose
09:41:16	Araq	all that it would accomplish is slower compile-times
09:41:58	Araq	in contrast to that Go is actually an interesting compiler target, we offload the task of having to write a multi-threaded GC to somebody else
09:42:52	FromDiscord_	<sealmove> Are you sure it's impossible to generate lifetime annotations without the source language having it?
09:42:55	*	al_ joined #nim
09:43:19	Araq	I'm such it's possible to generate it for Nim for many many cases.
09:43:36	Araq	but I don't know if the last missing 20% would kill it
09:43:48	Araq	*sure
09:47:10	FromDiscord_	<sealmove> Hmm, little confused about Nim->Go. Nim can use Go's GC, no?
09:47:26	Araq	true but I don't know how it works :P
09:48:51	Araq	compiling to X usually means you get good interop with X if you do it properly, so you could easily use Go libs
09:50:15	*	CcxWrk quit (Remote host closed the connection)
09:50:31	FromGitter	<xmonader> @Araq you mean go assembler itself right?
09:51:58	FromGitter	<alehander42> sealmove but that's the whole point: you have to add your lifetime anntoations so the compiler can know what to do
09:52:09	FromGitter	<alehander42> i am feeling that you're missing on the point of rust
09:52:56	FromGitter	<alehander42> if it was easy to generate somehow all of that invariants correctly based on normal code, somebody would've created such a language/type engine already
09:53:58	FromGitter	<alehander42> it's a bit like "why cant we compile ruby to C++ easily"
09:54:05	Araq	I wouldn't say that.
09:54:27	FromGitter	<alehander42> well that's how i see it, should've add IMHO
09:54:30	*	CcxWrk joined #nim
09:54:31	Araq	Before Rust came along not many could imagine it's possible at all.
09:54:48	Araq	so now that we know Rust can do it, people can think about inferring more and more lifetimes
09:54:52	FromGitter	<alehander42> ok, but it still sounds hard to infer them
09:55:09	FromGitter	<alehander42> oh yes, but my point is that this is the *hard* problem
09:55:20	Araq	I think the bigger problem is that even if you infer them all
09:55:25	FromGitter	<alehander42> not something that would be easy to just get for free while compiling to rus
09:55:31	Araq	in the end it's more restrictive than not having them.
09:55:53	*	xet7 quit (Read error: Connection reset by peer)
09:56:11	FromGitter	<alehander42> otherwise sealmove you're right that more attention should be given to ada spark and other similar systems, but yeah the marketing here works
09:56:43	FromDiscord_	<sealmove> It wasn't me that mentioned Ada/spark
09:56:49	FromGitter	<alehander42> well yes, that's why an ability to set some of them explicitly and write unsafe blocks somewhere for 1 5% of your code
09:57:01	FromGitter	<alehander42> seems still reasonable
09:57:03	FromGitter	<alehander42> oh sorry
09:57:36	*	xet7 joined #nim
09:57:49	FromDiscord_	<sealmove> But I get your point. If you can infer lifetimes, why even bother compiling to Rust...
09:58:07	FromGitter	<alehander42> honestly i can imagine something like an interactive(refactor-like) system which suggests different restrictions and the programmer refining using it
09:58:15	FromGitter	<alehander42> but i guess this will backfire a lot
09:58:20	FromGitter	<alehander42> sealmove exactly!
09:58:27	Araq	Haskell is pretty good at type inference
09:58:40	*	nc-x joined #nim
09:58:46	Araq	still working with it is completely different than working with Python.
09:58:49	nc-x	doesn't scopes infer lifetimes or something like that
09:58:56	nc-x	?
09:59:12	stefanos82	Araq: I hope one day to have "The Design and Evolution of Nim Language" into my hands signed by you. I have so many things to learn about decision-making during the design and implementation of a programming language.
09:59:16	FromGitter	<alehander42> yeah they had that "sidebands" idea, but not sure if related
09:59:38	FromDiscord_	<sealmove> There is another thing though. Rust has strong macros. The syntax is a nightmare but it's possible to write a language in Rust macros.
09:59:39	Araq	and I'm not saying that Haskell is bad, I'm saying people don't move from Python to Haskell just because it can infer "all the important things"
10:00:13	nc-x	Scopes: Compile-time resource management using view propagation, an annotation-free variation of borrow checking.
10:00:14	FromGitter	<alehander42> Araq, the other thing i wonder is if it's easy to have a language both good for low system programming with lifetime-like infereence moves etc and for high level gc-based stuff
10:00:43	FromGitter	<alehander42> basically nim is going in this path, and i am still not sure how well would different libs/parts of the lang coexist
10:00:50	Araq	stefanos82, that would be awesome indeed :-)
10:00:57	nc-x	Anyways how long for the release?
10:01:29	Araq	nc-x, we have Scope's author here in #nim, @lritter iirc
10:02:39	narimiran	nc-x: by my calculation, it will take around 3 more hours for nightlies to test everything
10:02:39	nc-x	yes i remember he gave some explanation regarding the view progagation recently
10:02:48	nc-x	narimiran: okay
10:03:22	Araq	the problem with lifetimes is that they fail to model a heap.
10:03:33	nc-x	but imo the nightlies travis script needs to be split up to use different CI services. currently it takes too long for them to build.
10:03:58	FromGitter	<alehander42> Araq, is there an alternative mechanism that can model them
10:04:06	nc-x	or make travis somehow prioritize win/linux/mac to be built first bcos they are most common platforms
10:04:07	narimiran	nc-x: there's some work done by leorize[m] to use azure pipelines, which should speed up things quite a bit
10:04:14	Araq	haven't found any, I keep looking, alehander
10:04:33	dom96	Nim->Go? Please no
10:04:34	nc-x	or if nim repo is green, then nightlies should not run tests but directly build
10:04:50	dom96	It's tough enough to convince Go fans of Nim at my workplace as it is
10:04:55	FromGitter	<alehander42> i also think nim should have its own good concurrency story
10:05:04	FromGitter	<alehander42> and no need of go tooling/tech dependence
10:05:05	nc-x	narimiran: yeah. azure offers parallel builds up to 10 iirc, so it should be a huge improvement.
10:05:20	narimiran	nc-x: exactly. plus it is faster in general
10:05:32	Araq	my idea is instead of using labels as lifetimes we use more general locations to describe what's going on
10:05:32	dom96	alehander42: Nim does have a good concurrency story
10:05:34	dom96	alehander42: async/await is concurrency
10:05:38	FromGitter	<alehander42> Araq, but what's the biggest weakness of lifetimes for heaps: what is a usecase where they fail a lot
10:06:33	FromGitter	<alehander42> @dom96 yes, that's the one (but still i remember it had some leaks etc , or this was a jester specific issu)
10:06:43	leorize[m]	narimiran, nc-x: well, seeing how supporting the entire CI set requires at least 4 jobs. This would effectively means only 2.5 pr/commit can be built at one time (haven't tried)
10:07:00	narimiran	"only"
10:07:00	dom96	alehander42: Those are pretty much fixed with my latest PR
10:07:08	FromGitter	<alehander42> ah great
10:07:12	nc-x	IMO, we should not commit to using a single CI service
10:07:25	Araq	you have an array of items, you delete one entry inside it, not all of them. Ensure that afterwards you don't access this element any longer.
10:07:30	nc-x	lets have travis for linux, appveyor for windows, azure for mac or something like that
10:07:33	dom96	Which isn't in 0.20.0, but 0.20.2 will surely not be far behind
10:07:46	Araq	you can count on 0.20.2 :-/
10:07:53	nc-x	then it will be faster. plus if any one of them go down, then things won't stop here.
10:08:06	federico3	dom96: I wouldn't call it good
10:08:18	narimiran	nc-x: based on the speed differences, maybe azure can do half of everything, and travis and appveyor do 1/4 each
10:08:23	Araq	nc-x, I completely agree but switching the CIs is not something we do before a release
10:08:25	FromGitter	<alehander42> Araq isn't this more about being hard to analyze collections
10:08:36	FromGitter	<alehander42> but that's related to heap overally probably yeah
10:08:44	leorize[m]	get tid of appveyor all together :p that's the biggest bottleneck
10:08:46	nc-x	Araq, narimiran: yeah
10:08:51	dom96	federico3: Why not?
10:08:58	leorize[m]	one job at a time is terrible
10:09:44	narimiran	leorize[m]: yeah, ok, we can polish the details (azure 2/3, travis 1/3?) so we can have more concurrent/parallel tests
10:10:51	leorize[m]	sure, and if I can implement job splitting, it's possible to crunch through the entire Nim's test suite in 10m or less
10:12:56	federico3	so much pain for CI
10:13:32	dom96	alehander42: The main cause of leaks (I hope) is stuck futures, I was able to confirm it in a game I'm developing so this should extend to other projects as well
10:14:02	*	nc-x quit (Quit: Page closed)
10:17:41	*	leorize joined #nim
10:22:05	narimiran	while we wait for new nim, here is new crystal release: https://crystal-lang.org/2019/06/05/crystal-0.29.0-released.html
10:22:46	narimiran	"There were 116 commits since 0.28.0 by 24 contributors." vs. >1000 commits for Nim new version ---> we should release more often :)
10:27:24	FromGitter	<xmonader> but they have a much smaller core language @narimiran
10:27:45	narimiran	i wasn't 100% serious ;)
10:28:41	FromGitter	<xmonader> xD :D I like what you did so far on the docs, can you maybe start some work on `effective nim` like effective go somehow?
10:28:59	narimiran	thanks :)
10:29:39	narimiran	i have an idea to make the sequel to 'Nim Basics', but i haven't started working on it and i don't know when/if i will
10:29:52	leorize	when can I expect nim 0.20? :p
10:30:00	narimiran	leorize: today :)
10:30:36	leorize	I should merge the nim 0.19.6 pr to haikuports then
10:30:41	*	laaron quit (Quit: ZNC 1.7.1 - https://znc.in)
10:30:54	*	laaron joined #nim
10:31:20	Araq	what? are you saying Haiku is always one version behind?
10:31:37	leorize	I'm busy :p
10:31:52	stefanos82	question: is there a particular reason "megatest" use a single core during execution?
10:31:57	narimiran	at least it is not like bleeding-edge Archlinux, which as of today will be 5 versions behind
10:31:58	leorize	haven't got the time to test it
10:32:15	leorize	note: rust is up-to-date there :P
10:34:33	krux02	stefanos82, well the compiler always uses just one core, The parallelism comes when multiple instances of the C compiler are invoked to compile the generated C files.
10:34:59	krux02	parallelizing a compiler is hard. calling a compiler multiple times in parallel is easy.
10:35:15	stefanos82	krux02: I thought megatest would also execute concurrent / parallel tests within it
10:35:44	stefanos82	that's the impression I originally had to be honest with you
10:35:51	krux02	the point of megatest is to join a lot of test to reduce the overhead, not to make anything parallel.
10:36:32	krux02	calling the compiler on a small file requires to process system with all its dependencies, this is often multiple times the size of the testfile itself.
10:37:00	krux02	when you join all the test files, you still need to process system.nim (and other dependencies) just once
10:37:34	krux02	so instead of compiling system.nim 500 times it now just compiles it one time.
10:38:09	stefanos82	krux02: what I mean is that I thought part of inner tests were also testing concurrency and parallelism, much like we have under tests for some cases
10:38:25	krux02	stefanos82, btw, are you stefantalpalaru on github?
10:38:31	stefanos82	nope
10:38:35	krux02	ok
10:39:26	krux02	with parallelism I have one guide I alwas follow. First check if you can optimize anything else. Then do that.
10:39:38	krux02	Parallelism creates complexity and is error prone.
10:39:52	krux02	and it doesn't promise too much scale.
10:40:29	krux02	a theoretical maximum of 8 times faster on 8 cores, but only when there is absolutely no overhead.
10:41:08	Araq	it's not that error prone, parallelism is quite easy, concurrency is the hard thing
10:41:45	stefanos82	often people confuse these two terms
10:41:59	Araq	however, it can still be tedious
10:42:48	stefanos82	in my opinion, to remedy this situation with concurrency demands a complete redesign of CPUs
10:42:56	krux02	I don't distinguish them that much as other people. But parallelism is about performance and concurrency is about control flow
10:43:13	krux02	stefanos82, we already have GPUs
10:43:19	stefanos82	krux02: true
10:43:45	krux02	it just needs to be simpler to program for GPUs.
10:44:25	krux02	then we are at my Rendering DSL again.
10:47:28	stefanos82	too bad I don't have an actual use of Nim's important usage. I got stuck with PHP and WordPress...don't ask! -_-
10:47:44	krux02	nim used to compile to PHP
10:47:49	*	dddddd joined #nim
10:48:00	krux02	(as an option)
10:48:49	stefanos82	to me Nim is an important language that I need to keep an eye on it
10:50:03	stefanos82	what V language claims to achieve, is what I wanted Nim to do
10:50:05	*	seni joined #nim
10:50:20	stefanos82	but I think Nim is getting there already
10:52:28	leorize[m]	Araq: has the =move went live or I'd have to wait for 0.21?
10:52:45	stefanos82	my dream would be to see Nim's execution time reaching microseconds and milliseconds at most
10:52:49	stefanos82	that would be mind blowing!
11:01:40	*	nc-x joined #nim
11:01:58	nc-x	leorize: wait for 0.21
11:02:50	stefanos82	guys, is this behavior normal? http://paste.debian.net/1086533/
11:03:00	stefanos82	I tested the example from the official website
11:03:32	stefanos82	my version is git hash: 8443392db5babc4b08db7f021e6dcd6cfa26a969
11:03:44	ehmry	stefanos82: commas?
11:04:28	nc-x	stefanos82: you forgot the comma
11:04:54	nc-x	between the two Person(..) objects
11:05:15	nc-x	inside people
11:06:49	stefanos82	lol darn it
11:06:52	stefanos82	good catch folks
11:07:11	*	nc-x quit (Quit: Page closed)
11:08:57	Araq	leorize[m], no =move for 0.20, sorry
11:09:07	Araq	but we have a good spec now... *cough*
11:14:49	leorize[m]	shashlick: I've just looked at nightlies, you can skip brew update for shorter build time
11:18:15	*	deech joined #nim
11:35:50	*	nsf quit (Quit: WeeChat 2.4)
11:42:18	lqdev[m]	shashlick: is nimterop capable of cImporting single-header libraries?
11:45:32	*	manniq31 joined #nim
11:46:21	*	manniq31 quit (Client Quit)
11:57:24	*	Kaivo quit (Quit: WeeChat 2.4)
11:59:31	*	AndChat|624225 joined #nim
12:00:25	FromGitter	<kaushalmodi> From a Twitter conversation, someone is wishing that Nim had ADT.
12:00:53	FromGitter	<kaushalmodi> What's ADT? Abstract Data Type?
12:01:12	*	NimBot joined #nim
12:01:37	FromGitter	<kaushalmodi> Wouldn't JsonNode, etc be abstract data type? I don't get that buzz word.
12:02:08	*	envoyt quit (Ping timeout: 248 seconds)
12:02:11	FromGitter	<xmonader> @kaushalmodi I do wish so too!
12:02:41	FromGitter	<kaushalmodi> Coming from a non CS background, what's an ADT all about?
12:03:15	FromGitter	<kaushalmodi> I read online that an ADT has that type specific functions.. but we have that almost everywhere in Nim.
12:03:16	FromGitter	<xmonader> it's algebraic not abstract
12:03:34	*	nc-x joined #nim
12:03:58	FromGitter	<xmonader> roughly it's about having some types like ⏎ type PaymentType = CreditCard Number | PayOnDelivery
12:04:33	FromGitter	<kaushalmodi> Ok, I ended up on https://en.m.wikipedia.org/wiki/Abstract_data_type
12:04:55	nc-x	narimiran, Araq: why does the blog post say nothing about the newruntime - its current state, new spec, and the future plans? imo it should be part of it. It is of interest to a lot of people.
12:05:40	FromGitter	<xmonader> https://en.m.wikipedia.org/wiki/Algebraic_data_type u need this one
12:05:59	FromGitter	<xmonader> I know variant exists in nim, but it's just ugly to use.
12:06:50	FromGitter	<kaushalmodi> Would concept fit the bill (not that I have used it): https://nim-lang.github.io/Nim/manual_experimental.html#concepts
12:07:19	FromGitter	<kaushalmodi> From the looks, it seems like it constraints the value of a data type using some user provided logic.
12:09:19	FromGitter	<kaushalmodi> Well, if you or anyone wants to further this discussion on ADT, here's the thread on Twitter: https://twitter.com/mmstick/status/1136439753971134464?s=09
12:10:43	FromGitter	<xmonader> concepts looks like interfaces, or typeclasses in Haskell, the example for comparable is like `Eq`
12:11:37	*	nc-x quit (Ping timeout: 256 seconds)
12:12:00	leorize	kaushalmodi: iirc gara has more features than patty, but I'm not sure since I never saw a need for pattern matching :p
12:14:16	FromGitter	<xmonader> in short, you can describe sum types using enums, do enums take params? ⏎ ⏎ and product types can be represented with tuples or records or at least that's how i do it
12:15:57	leorize	afaict, these types are just your normal case object without the kind field
12:16:55	FromGitter	<xmonader> @kaushalmodi https://github.com/xmonader/nim-redisparser/blob/master/src/redisparser.nim#L6 from line 6 to 54 in haskell u would do something like this https://github.com/informatikr/hedis/blob/master/src/Database/Redis/Types.hs#L52
12:17:45	FromGitter	<kaushalmodi> leorize: neither have I needed pattern matching. I've heard of gara, will mention that too.
12:18:14	leorize	@xmonader: with a bit of macro magic that kind of syntax is also possible with Nim
12:18:30	*	lritter joined #nim
12:20:37	FromGitter	<xmonader> it will make our lives easier, i faced the same thing for my beconde library so instead of doing ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5cf905156f530d3b6147a01c]
12:21:10	leorize	you can just not qualify the enums...
12:21:41	FromGitter	<xmonader> are you talking about pure?
12:21:52	leorize	well but doesn't appear to be too hard to emulate that syntax
12:22:11	leorize	@xmonader: you never have to qualify enums...
12:22:44	FromGitter	<xmonader> well, leorize then after than you will have boilerplate for $ , == and hash while in haskell u can always do deriving Show, Eq for such
12:24:28	leorize	do you feel like writing a macro? :) I'm pretty sure this is possible with some macro magic
12:25:11	FromGitter	<xmonader> ah I know, because I didn't settle on enum style yet, in pascal u usually have some type abbrev in the beginning and in some languages that forces qualified name like python or C# .. i like to use fullnames most of the time
12:25:32	FromGitter	<xmonader> nah, too many edge cases and my macro powers aren't that good in nim
12:26:23	leorize	in Nim use abbrev in the beginning
12:27:15	FromGitter	<xmonader> or use {.pure.} ?
12:28:40	leorize	pure will be deprecated IIRC
12:30:03	leorize	there'll be a kind of qualification required if collide for enums like the one for modules iirc (not sure if Araq implemented that yet)
12:32:46	FromGitter	<xmonader> patty sounds to have that macro @leorize !! yay ⏎ https://github.com/andreaferretti/patty/blob/master/README.md#constructing-variant-objects
12:33:44	leorize	told you it's possible :P
12:35:03	*	deech quit (Ping timeout: 245 seconds)
12:44:34	*	zahary joined #nim
12:51:33	FromGitter	<arnetheduck> patty would work well with https://github.com/arnetheduck/nim-result :)
12:52:00	*	xet7_ joined #nim
12:52:21	*	xet7_ quit (Remote host closed the connection)
12:52:45	*	xet7_ joined #nim
12:55:03	*	xet7 quit (Ping timeout: 245 seconds)
12:55:19	*	xet7_ is now known as xet7
12:56:58	*	AndChat|624225 quit (Ping timeout: 258 seconds)
12:58:08	FromGitter	<xmonader> i like that! @arnetheduck
12:59:15	FromGitter	<xmonader> I know Araq hates options/eithers, and I'm sure the stdlib won't utilize that in favor for exeptions
12:59:43	*	Snircle joined #nim
13:00:20	FromGitter	<alehander42> are there any apps comparing exceptions/results for the same code
13:00:29	FromGitter	<alehander42> this woudl really help to compare, otherwise good lib
13:00:59	*	brainproxy joined #nim
13:01:49	Araq	'hate' is a strong word, but if you try to sell them as an "alternative" to exceptions than I still cannot take you seriously
13:07:00	FromGitter	<xmonader> Well, many others design their libs with that as alternative to exceptions. And everyone is taking them seriously :)
13:08:41	leorize	certainly not nim users, most libs are designed with exceptions :)
13:08:55	Araq	so where are these new languages where '+' returns an Option[int]?
13:10:05	Araq	all they do is to map IO errors into Either monads and delude themselves into thinking that the design "scales". It doesn't which is why array index out of bounds traps in Rust, it doesn't return an Option[ElementType]
13:10:57	livcd	whoa a i was just searching for a very generic thing and google gave me a github repo for a project in nim
13:11:19	leorize	you've been using google for too long :p
13:11:27	FromGitter	<xmonader> @Araq won't that be fixed with dependable types? so instead of array[5] you have type ArrayOf5 ?
13:11:34	*	envoyt joined #nim
13:11:38	FromGitter	<arnetheduck> @alehander42 https://en.cppreference.com/w/cpp/experimental/fs/copy - have your pick :)
13:11:51	livcd	leorize: i was searching for "text expander for windows"
13:12:13	livcd	and got https://github.com/ajusa/sanictype as a first hit
13:12:55	leorize	yea, you've been using google for too long :p
13:13:05	leorize	https://www.startpage.com/do/metasearch.pl?query=text%20expander%20for%20windows
13:13:18	leorize	^ that's what normal people see
13:13:35	leorize	I mean people that don't use google as much :p
13:14:22	Araq	personalized searches rule :-)
13:15:33	FromGitter	<xmonader> also another question, does nim `forces` you to handle exceptions in the calling blocks?
13:15:52	Araq	use .raises: [] and the compiler does force you to handle it
13:16:41	leorize	Araq: does that work for `block:` statements?
13:17:00	FromGitter	<xmonader> It sounds like a sane default maybe .noraises. should be provided instead
13:17:16	Araq	that has been proposed indeed.
13:17:50	*	apodo joined #nim
13:19:08	FromGitter	<xmonader> much safer :)
13:19:19	Araq	by what measure?
13:20:09	Araq	forcing me to ignore errors in order to comply to .noraises is not 'safer'
13:21:42	FromGitter	<xmonader> It's the programmer choice to ignore it, but what i expect from the language is to tell me there's something i need to handle (which is only activated with raises)
13:21:44	livcd	what happened to ryukoo
13:22:00	narimiran	livcd: that's what i'm wondering too
13:23:22	livcd	well he is active on gitlab
13:23:40	FromGitter	<xmonader> the case with `Result` type is when you match against it right there you will handle the code for Error (or ignore it) but with exceptions it may go unnoticed unless you control the library you're using to `force` raises.
13:24:07	*	Araq sighs
13:24:16	Araq	here is what happens in reality:
13:24:25	Araq	you forget to catch the exception
13:24:39	Araq	the program dies a bit too early with a quite instructive error message
13:25:20	Araq	here is the gain of the idea: code size reduction by a factor of 2.
13:25:27	Araq	easily.
13:25:38	Araq	cost vs benefit analysis, it's a thing.
13:25:47	disruptek	seems a little unfair to claim that .raises. is broken on Monday and then claim that .raises. solves anything on Tuesday.
13:26:42	Araq	disruptek, .raises does solve the problem that programmers think they have, but it doesn't solve the *real* problem
13:27:05	livcd	narimiran: https://gitlab.com/ryukoposting/northstar
13:27:30	disruptek	i guess i'm the one that claims that .raises. is broken, and you claim that it works fine, it's just semantically off-point, right?
13:27:31	narimiran	livcd: ok, still using nim, i was afraid he maybe stopped....
13:28:34	Araq	disruptek, I don't like exceptions but so far I'm still looking for convincing arguments against them
13:29:19	disruptek	it does seem that there should be a better way. the whole concept lacks a certain elegance, doesn't it?
13:29:24	*	theelous3_ joined #nim
13:29:44	Araq	"blah, blah, hidden control flow" doesn't cut it when the proposed alternative introduce just as many different control flow paths
13:30:13	Araq	but this time "hidden" in all the boilerplate
13:30:46	Araq	it's like getters and setters in Java, it's not that you have write them all the time, you also need to actively ignore them in every code base
13:32:48	disruptek	exception handling boilerplate isn't the deficiency, i think -- it conveys the intent of the programmer both to the compiler and other programmers. that part isn't such a wart, in my opinion. it's the almost arbitrary control flow that seems most painful. and it's much harder to reason about, both for the compiler and the programmer.
13:37:33	leorize	the last thing I ever want to do is to handle IOError from `echo`
13:38:12	dsuch	stefanos82: Having just learned of its existence by following a link of yours, is that honestly the whole of it? https://vlang.io/docs
13:38:35	stefanos82	I guess dsuch
13:38:45	stefanos82	we are waiting for the official release by the end of June
13:39:39	Araq	disruptek, boilerplate has costs too, more work to refactor, more code to read, more code that can have bugs
13:39:59	disruptek	Araq: and what's bad about Go's version, if you can remind me succinctly?
13:40:49	dsuch	stefanos82: Is that a language by Microsoft, Google or Apple? I just see a bunch of basic todos all throughout that page and I wonder how come this repository has 3k+ stars
13:41:03	disruptek	dsuch: outrageous claims.
13:41:12	leorize	vlang seems to pack with so many features it's too good to be true
13:41:18	narimiran	dsuch: hype-based development
13:41:31	stefanos82	dsuch: no, it's a one man's project
13:41:31	Araq	can we have the vlang discussions in #nim-offtopic please?
13:42:13	dsuch	Araq: I did not realize such a channel existed and that it would be preferred
13:42:36	dsuch	Thanks anyway disruptek et al.
13:42:37	Araq	disruptek, what's good about it? Go has panic/recover which is an exception system
13:42:55	*	dsuch left #nim (#nim)
13:43:12	Araq	it merely uses it less frequently. which means that you can write your 'close' not in a 'defer' environment... er.. ooops
13:43:13	disruptek	the error passing through multipart return values.
13:44:38	Araq	for me an exception system models that every function can fail. Yes, you don't like that because math is different. But programmnig is not math, in programming even integer additions can fail.
13:45:08	disruptek	i agree.
13:45:36	*	envoyt quit (Ping timeout: 248 seconds)
13:46:02	lqdev[m]	we need to make a language that's safe no matter what you try to do: `(2 + 2).unwrap("Couldn't add 2 + 2")`
13:46:02	*	envoyt joined #nim
13:46:33	Araq	we need to make the .unwrap("good error message") be the default
13:46:33	disruptek	so why not just have every function measure success or failure, and maybe, (just maybe) have them flag the nature of the failure.
13:47:55	lqdev[m]	for me the whole idea of safety through constant `.unwrap()` calls everywhere in your codebase is far from preferable
13:48:16	lqdev[m]	it just adds a ton of noise and makes code unreadable
13:48:37	disruptek	it's a logical construct, not a literal one. just imagine that it's there.
13:49:10	FromGitter	<alehander42> but i am not sure why is it improtant to deal with a+b
13:49:15	FromGitter	<alehander42> this is an exceptional case
13:49:21	FromGitter	<alehander42> thats where exceptions would shine
13:49:26	federico3	I'd love a go2nim translator
13:49:27	FromGitter	<alehander42> and you can still have error enums
13:49:31	FromGitter	<alehander42> for the more usual cases
13:49:34	lqdev[m]	exceptions are a good tradeoff between readability and safety, imo
13:49:39	federico3	there's plenty of stuff that could be converted
13:49:40	FromGitter	<alehander42> basically use both
13:50:08	Araq	alehander42: my 'echo' statements can fail. rarely. it's an exceptional case
13:50:14	FromGitter	<alehander42> ok
13:50:39	FromGitter	<alehander42> so, again, exceptions can deal with such cases which are rarely handled: mostly on the top of a subsystem
13:50:51	FromGitter	<alehander42> and error enums with the other ones imo
13:50:59	disruptek	the point that integer addition may fail is unassailable and useful as a measure of such a system.
13:51:13	FromGitter	<alehander42> (including bubbling errors automatically so you can still raise them thru the callstack with no try/catch)
13:51:22	disruptek	you make it sound as though you support optional error handling.
13:51:25	FromGitter	<alehander42> (e.g. callwhichcanerr?(e))
13:53:48	*	zahary quit (Ping timeout: 245 seconds)
13:55:40	*	PMunch quit (Remote host closed the connection)
13:55:57	*	zahary joined #nim
13:58:01	*	aredirect joined #nim
14:00:27	*	zahary quit (Ping timeout: 248 seconds)
14:02:55	*	brainproxy quit (Quit: WeeChat 2.4)
14:07:36	shashlick	@federico3 I could adapt nimterop to any language tree-sitter supports but Nim needs to add an importgo equivalent
14:08:55	shashlick	I know I missed the CI conversation but Travis for nightlies shares the same 5 build slots as our main devel CI along with every other repo in nim-lang
14:09:49	shashlick	No doubt it is slower but moving should be easier since most of the code is in sh files rather than yml
14:10:40	shashlick	Although arm builds will require more work to ensure qemu is setup correctly
14:15:11	leorize	shashlick: maybe you can offload a bit to circleci?
14:16:02	leorize	with 36 cores, that thing crunch through the entire Nim CI procedure for Linux in 18mins
14:16:13	leorize	or 20, I don't remember :p
14:19:14	shashlick	Should be doable but reason I've not optimized nightlies is that it happens every day once and doesn't gate anyone
14:19:34	shashlick	Of course except when you want to make a release!
14:28:00	*	aredirect quit (Quit: Leaving)
14:31:04	*	envoyt quit (Ping timeout: 272 seconds)
14:31:27	*	envoyt joined #nim
14:32:06	*	ajsopry joined #nim
14:33:55	*	ajsopry left #nim (#nim)
14:48:49	*	al_ quit (Quit: al_)
14:58:52	livcd	where's yardanico ?
14:59:17	*	envoyt quit (Ping timeout: 252 seconds)
14:59:52	FromGitter	<kaushalmodi> narimiran: there's a typo in the v0.20.0 blog post
15:00:10	FromGitter	<kaushalmodi> ```# v0.20: ⏎ const b = uint16(-1) ⏎ # Error: -1 can't be converted to uint8``` [https://gitter.im/nim-lang/Nim?at=5cf92a7af3a60a79a4641aba]
15:00:12	narimiran	kaushalmodi: where? give me the context
15:00:26	FromGitter	<kaushalmodi> the conversion is to uint16 but error says uint8
15:00:35	narimiran	right, nice catch
15:01:12	leorize	well, 0.20 coming anytime soon? :p
15:01:31	leorize	finally nim.nvim will work on stable
15:01:56	FromGitter	<kaushalmodi> leorize: it's already announced
15:02:10	FromGitter	<kaushalmodi> and the repo is tagged
15:02:20	*	apodo left #nim (#nim)
15:02:23	Araq	hmm 388 bugfixes
15:02:48	Araq	quite an achievement
15:03:29	Araq	ah, it's only 387
15:03:55	leorize	woo, got #7 on the commit chart
15:05:42	FromGitter	<kaushalmodi> heh, I am surprised to be even there.. number 27 :P
15:07:27	narimiran	kaushalmodi: fixed
15:07:30	leorize	nim.nvim readme updated :D
15:08:00	narimiran	celebration time
15:08:12	narimiran	(until you fuckers report the first regression)
15:08:23	FromGitter	<kaushalmodi> narimiran: cool
15:08:41	FromGitter	<kaushalmodi> this has been bugging me: https://github.com/nim-lang/Nim/issues/10800
15:09:00	FromGitter	<kaushalmodi> as I import `os` in my `config.nims` I cannot run `koch docs` on my machine
15:09:07	narimiran	so literally i cannot have one minute of celebration? :(
15:09:13	FromGitter	<kaushalmodi> hehe
15:09:19	FromGitter	<kaushalmodi> it's not a 0.20 regression
15:09:34	dom96	Araq: Just call it 400 :)
15:09:45	FromGitter	<kaushalmodi> it's been there.. I just tested if it still errored on 0.20.0 too.. and it does
15:10:11	*	narimiran la-la-la don't hear you
15:10:43	FromGitter	<kaushalmodi> :) today's for celebration!
15:10:53	leorize	on HN yet? :p
15:11:16	narimiran	leorize: tomorrow morning would be a better timing i think, but whatever
15:11:32	FromGitter	<kaushalmodi> it's not even on twitter yet
15:11:39	federico3	blog post?
15:11:49	leorize	blog post is up
15:11:49	FromGitter	<kaushalmodi> this announcement should be a pinned tweet IMO
15:11:57	FromGitter	<kaushalmodi> the current pinned tweet is too old
15:12:22	Zevv	congrats an happy 0.20 day everybody
15:12:45	*	jasper_ joined #nim
15:12:56	*	envoyt joined #nim
15:13:41	narimiran	Zevv: thanks, cheers! :)
15:13:51	dom96	Tweeted
15:13:54	shashlick	congratulations!
15:13:57	Zevv	is the "no breaking guarentee" for the core language only, or also for the stdlib?
15:14:08	dom96	Zevv read on :)
15:14:08	federico3	the naming convention of the xz tarball has changed
15:14:25	FromGitter	<xmonader> congrats on the release and all the hard work
15:14:27	federico3	this time it's nim_0.20.0-linux_x64
15:14:29	Zevv	dom96: yeah right sorry about that :)
15:14:38	dom96	pinned tweet too
15:15:03	narimiran	federico3: that's binary, source should be the same as before, no?
15:15:45	leorize	read 1/3 of the page and the `shr` change is nowhere to be seen :p
15:16:01	FromGitter	<kaushalmodi> shouldn't the version number be mentioned somewhere on the install pages like: https://nim-lang.org/install_unix.html
15:16:05	Zevv	leorize: read on :)
15:16:09	dom96	leorize: I even bolded it for you
15:16:35	narimiran	dom96: i think the bold was removed at the end, sorry :)
15:16:40	dom96	:'(
15:16:47	narimiran	but it is the first thing in the changelog
15:16:50	FromGitter	<kaushalmodi> but it's the first point in changelog
15:17:01	leorize	can an we have example as well? so that it'd stand out a bit
15:17:46	Zevv	given the number of fixes 0.19 must have been severely broken :)
15:17:47	narimiran	i guess we can copy the examples from https://nim-lang.org/docs/system.html#shr%2Cint%2CSomeInteger
15:20:54	leorize	narimiran: why does the doc for shr still recommends `ashr`?
15:20:57	FromGitter	<xmonader> i love this one ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5cf92f59e41fe15e752198cd]
15:22:52	*	PMunch joined #nim
15:24:34	PMunch	Congrats on 0.20.0
15:30:00	leorize	!eval echo NimVersion
15:30:04	NimBot	0.19.6
15:30:09	leorize	PMunch: ^
15:34:02	PMunch	Ooh, good point
15:35:16	PMunch	https://hub.docker.com/r/nimlang/nim/tags
15:35:29	PMunch	Hmm, this is still on 0.19.6, which is what the playground uses
15:35:42	leorize	I thought those are unofficial?
15:36:21	PMunch	Yeah I think so
15:36:32	PMunch	Maybe I'll just roll my own..
15:38:08	leorize	PMunch: https://github.com/moigagoo/nimage
15:41:15	PMunch	Hmm, yeah I really should make a staged one
15:41:32	PMunch	So it builds it in one container then copies only the finished stuff into the final container
15:42:09	*	nc-x joined #nim
15:42:22	nc-x	@KaushalModi https://github.com/nim-lang/Nim/pull/11418
15:42:35	leorize	https://old.reddit.com/r/programming/comments/bxhw0c/nim_version_0200_is_released/
15:42:38	leorize	for upvoting :p
15:42:55	nc-x	Congrats on 0.20 everyone!!!
15:43:18	nc-x	somebody also needs to post on nim forum regarding the release.
15:43:33	leorize	you have an account, right? :p
15:43:44	nc-x	nope
15:43:49	nc-x	;D
15:48:34	FromGitter	<kaushalmodi> nc-x: thanks!!!
15:49:39	federico3	https://lobste.rs/s/7h70w1/nim_0_20_0_has_been_released
15:51:10	dom96	hah, funny, Crystal release just yesterday as well
15:51:33	dom96	can we get more upvotes? ;)
15:51:43	FromGitter	<xmonader> and clojure too :D
15:51:49	narimiran	yeah, and ocaml is bringing new release too
15:52:04	leorize	crystal website looks rather nice though
15:52:18	dom96	I prefer Nim's, but I am pretty biased :P
15:52:23	FromGitter	<xmonader> can someone comment on the sshr operator history question on reddit?
15:52:34	leorize	there's already `ashr`
15:52:48	leorize	the shr change is to conform with other languages, and to get rid of ashr
15:52:50	dom96	looks like def- beat me to HN heh
15:52:58	dom96	Thank you for including "1.0 RC" in there
15:53:18	leorize	xmonader: someone asked the same on the forum :p see the rant about shr
15:53:53	FromGitter	<xmonader> 1) 0 RC 1 is soo clickbaity i wish I included it in the title
15:54:19	dom96	it's not click bait when it's true :)
15:54:45	def-	dom96: seemed like an important point
15:54:51	FromGitter	<xmonader> true, maybe there's another word for what i meant it in English :)
15:56:22	leorize	@xmonader: https://forum.nim-lang.org/t/4891#30623
15:57:11	FromGitter	<xmonader> thank you!
15:57:43	leorize	https://news.ycombinator.com/item?id=20115532
15:59:00	dom96	Don
15:59:05	dom96	*Don't link to HN directly
15:59:20	leorize	why?
15:59:32	dom96	If people upvote by going straight to that link the thread will get banned
15:59:48	leorize	oh, never know about that, sorry
15:59:59	FromGitter	<xmonader> Really? wow. How the rust folks do it :D ؟
16:00:24	*	nc-x quit (Quit: Page closed)
16:00:25	FromGitter	<kaushalmodi> https://hn.algolia.com/?query=nim&sort=byPopularity&prefix&page=0&dateRange=pastWeek&type=story
16:00:43	leorize	they don't have to, the word "rust" is a HN buzzword now
16:01:32	dom96	Anyway, Nim is on the front page now so need for links ;)
16:01:57	FromGitter	<kaushalmodi> also "Medium sucks" is a buzz word on HN.. I must have come across 4-5 articles singing the same song and they are show up top on HN
16:02:10	FromGitter	<kaushalmodi> it's a fact that Medium sucks; people should now stop writing about that
16:02:40	*	envoyt quit (Ping timeout: 248 seconds)
16:02:53	PMunch	Woo, my nimsuggest changes have also been merged :) So now nim-lsp doesn't depend on devel any longer
16:03:33	dom96	kaushalmodi: lol
16:14:49	stefanos82	dom96: I didn't know about that thing
16:15:02	stefanos82	how come a link gets banned by accessing it directly?
16:16:03	narimiran	vote manipulation prevention
16:17:52	*	envoyt joined #nim
16:28:56	*	floppydh quit (Quit: WeeChat 2.4)
16:36:27	*	Kaivo joined #nim
16:43:16	FromGitter	<xmonader> ```code paste, see link``` ⏎ ⏎ I went through lots of trouble to generate dep free executable. [https://gitter.im/nim-lang/Nim?at=5cf942a44dfb4240048df2d5]
16:43:24	leorize	looks like Nim is doing well on HN :)
16:43:48	leorize	xmonader: your target was linux :p
16:44:15	leorize	which is just the worst target ever
16:44:18	*	terorie joined #nim
16:44:22	FromGitter	<xmonader> yeah and ssl was the dependency :D it's safe to assume most of developers are using linux
16:44:39	FromGitter	<xmonader> I never developed on other platform.
16:44:44	leorize	well, once we have native nim modules for everything (like go), we can relax and have deps free exec
16:45:22	leorize	for now, we're as deps free as any C program out there
16:45:49	FromGitter	<xmonader> but we are mentioning that as a nim feature, which aren't free really as what people expect
16:46:57	FromGitter	<jrfondren> did you expect to not have SSL as a dependency?
16:47:38	FromGitter	<xmonader> i expected easily built statically linked binary :)
16:47:42	FromGitter	<jrfondren> I read that as, "oh, it's not a scripting language".
16:47:55	FromDiscord_	<DeltaPHC> Has there been any desire to update the version of MinGW linked on the install_windows.html page? 6.3.0 is a tad old by now :P
16:47:58	*	krux02 quit (Remote host closed the connection)
16:48:02	leorize	Araq: ^
16:48:07	FromGitter	<jrfondren> static linking is possible, but it's not as easy as it could be yeah.
16:48:57	FromDiscord_	<DeltaPHC> Personally I use mingw-w64 8.2.0 with Nim and it works fine
16:49:21	FromGitter	<xmonader> @kaushalmodi where's the link to the experiment we had with static linking nim to ssl?
16:50:00	FromGitter	<xmonader> found it nvm https://github.com/kaushalmodi/hello_musl
16:50:24	FromGitter	<jrfondren> what's the musl for? to get DNS resolution without dynamic linking?
16:51:10	leorize	libc-independance
16:51:16	Araq	DeltaPHC: we can switch to 8.2.0 after extensive testing
16:51:23	Araq	but yeah, we should
16:51:33	*	nsf joined #nim
16:53:02	leorize	someone opened the style-insensitivity can of worms on HN :p
16:54:21	PMunch	Hmm, can you use https://nim-lang.org/docs/asyncdispatch.html#withTimeout%2CFuture[T]%2Cint with an async http request to give it a timeout?
16:54:33	FromGitter	<xmonader> no with async no.
16:55:00	leorize	no idea how the http timeout thing works
16:56:24	Araq	leorize, 'nim --styleCheck:warning' is a thing
16:56:25	FromGitter	<xmonader> @PMunch https://nim-lang.org/docs/asyncnet.html#recv%2CAsyncSocket%2Cint no overloaded version for recv with timeout
16:57:14	PMunch	xmonader, yeah it doesn't have a timeout there. But it returns a future so withTimeout from asyncdispatch should be able to do it right?
16:57:35	leorize	Araq: was it documented as a new feature for 0.20?
16:57:41	*	nc-x joined #nim
16:58:14	Araq	leorize, I think I forgot to mention it :P
16:58:32	leorize	narimiran: quick :p ^
16:59:01	narimiran	leorize: other, more important, worries currently
16:59:16	nc-x	I just remembered. Apart from newruntime, the hotcodereloading also did not have a mention in the blog post?
16:59:27	leorize	hcr is experimental rn
16:59:35	leorize	there aren't any docs yet iirc
17:00:38	FromGitter	<xmonader> @dom96 is that how it meant to be timeout async recv? because afaik there was some note it's not implemented somewhere
17:02:05	nc-x	leorize: I think they deserved a mention atleast near the top
17:02:29	nc-x	Also can anybody verify the nim stdlib docs don't have the search on mobile
17:02:49	leorize	newruntime can be held off a bit
17:02:55	nc-x	eg. here https://nim-lang.org/docs/algorithm.html
17:03:00	FromGitter	<jrfondren> @nc-x, they don't have the search on mobile.
17:03:07	leorize	Araq has just redesigned the thing before 0.20 :p
17:03:22	Araq	guys, can't you test this stuff *before* we release...
17:03:34	dom96	xmonader: not sure what you mean
17:03:45	Araq	always the same shit, .0 is crap, .2 is workable
17:04:19	nc-x	I think the docs can be updated without a new release
17:04:44	leorize	well not having search on mobile is not the end of the world
17:04:45	FromGitter	<xmonader> @dom96 asycnet.recv doesn't allow timeout param. is it meant to be used with `withTimeout` or something else?
17:05:06	Araq	nc-x, that's not the only problem, we also ship with a broken nimble
17:05:20	*	shadowbane quit (Remote host closed the connection)
17:05:41	*	Serenitor joined #nim
17:05:46	leorize	now that's not a nice thing
17:06:01	nc-x	broken nimble?
17:07:14	leorize	but nimble CI is green?
17:07:55	dom96	xmonader: yes, it's meant to be used with that
17:07:58	narimiran	leorize: don't ask....
17:08:17	Araq	it's our overly complicated testing + Nimble bundling setup :P
17:08:58	Araq	if only Nimble were part of Nim ... so much friction just because it's in its own repo, it's ridiculous
17:09:42	dom96	Nah, this is a simple case of: it should have been tested
17:10:13	dom96	anyone extracting one of the tarballs and testing would have caught this
17:10:25	Araq	it was tested. automatically. it also was tested that 'koch nimble' works.
17:10:37	Araq	we trust our automation :P
17:10:49	*	nc-x quit (Ping timeout: 256 seconds)
17:11:34	leorize	new CI pipelines: package nim then extract & test :p
17:15:44	*	fredrik92 joined #nim
17:16:43	*	couven92 quit (Disconnected by services)
17:16:56	*	fredrik92 is now known as couven92
17:19:29	Serenitor	I'm trying to get my code compiled on 0.20.0, last time it compiled was on 0.18.0
17:19:36	Serenitor	seems like there's quite a lot to fix
17:19:41	Serenitor	on thing is that I can no longer set a string to nil?
17:19:56	leorize	yea, looks like you missed out 0.19
17:21:02	narimiran	Serenitor: that was gone in 0.19 too. see 0.19 release notes on how to keep compiling that code
17:21:41	Serenitor	will do, thanks
17:25:36	*	couven92 quit (Read error: Connection reset by peer)
17:29:27	*	leorize quit (Remote host closed the connection)
17:29:53	*	leorize joined #nim
17:32:57	rayman22201	Congrats on the new release! I'm always sleeping when the fun stuff happens.
17:33:24	leorize	dom96: maybe it's time to update the topic?
17:45:17	*	couven92 joined #nim
17:45:18	*	couven92 quit (Client Quit)
17:56:31	FromGitter	<brentp> I am seeing a 2x slowdown in some of my code between 0.19.9 (5b1bc29e8ad61973fa543a146cf19506d4d7f72c)and 0.20.0 (7a670a6bc929ee497dfd119b4a502b5b1ee139b8) anyone else finding this?
17:56:55	leorize	brentp: -d:release now enables some checks
17:57:03	leorize	use -d:danger for no checks
17:57:06	narimiran	brentp: try with -d:danger
17:57:42	FromGitter	<brentp> aye. that does it. thanks @leorize and @narimiran
17:58:34	leorize	brentp: you can try tracking down the hotspot then use {.checks.} pragma to selectively disable runtime checks
17:58:44	FromGitter	<brentp> where can I see more about this? I can probably .push on some stuff in hot paths
17:59:57	leorize	https://nim-lang.org/docs/manual.html#pragmas-compilation-option-pragmas
18:00:31	narimiran	if anybody wants to join the discussion.... quite a lot comments here https://news.ycombinator.com/item?id=20115532
18:01:14	leorize	welp, dropped to bottom of first page
18:01:26	leorize	but US is waking up, so we might see some bounce back
18:14:08	PMunch	xmonader, dom96, since async recv is supposed to be used with withTimeout then maybe this section should be updated: https://nim-lang.org/docs/httpclient.html#timeouts
18:22:10	sealmove	guys I got an error (I know the cause) but got this error:
18:22:14	sealmove	main.nim(176, 6) Error: internal error: genLiteral(nkHiddenCallConv)
18:22:15	sealmove	No stack traceback available
18:22:17	sealmove	To create a stacktrace, rerun compilation with ./koch temp c <file>
18:22:59	sealmove	Is this normal?
18:23:37	sealmove	I mean for nim compiler to terminate like this. "internal error" sound severe :P
18:23:50	narimiran	well, it happens sometimes, if that's what you're asking. you might want to open an issue (or see if it is already open for a similar situation)
18:24:02	sealmove	yes that's why I am asking
18:30:05	leorize	narimiran: how's the nimble issue going?
18:30:25	narimiran	the fix is in devel
18:30:39	narimiran	but now i hear somebody having a different problem on windows....
18:31:24	narimiran	btw, in the mean time: those using `choosenim` can use `choosenim update devel`, `choosenim devel` to get a working nimble version
18:33:49	*	krux02 joined #nim
18:37:42	sealmove	I think it's a bug in `converter`
18:38:41	*	Serentty joined #nim
18:40:31	sealmove	https://termbin.com/af6v
18:41:14	sealmove	btw does this error mean *compiler crushed* ?
18:41:35	leorize	yea, an ICE means the compiler crashed
18:41:43	Serentty	Okay, so I saw that Nim is approaching stabilization, and I figured I would come on here to bikeshed something that has been bothering me for a while.
18:41:56	leorize	sealmove: you should file an issue on github
18:42:11	sealmove	yes, I'll do that
18:43:04	narimiran	Serentty: not today, not today.... :)
18:45:03	kungtotte	I hope it's about case insensitivity, we sure don't get enough discussions about that
18:45:27	Serentty	What's today?
18:45:36	narimiran	0.20 release day
18:46:46	Serentty	And I do think that this is fairly important, since it's about Unicode ergonomics, which is something that causes ripple effects leading to shoddy software when programming languages make irreversible decisions that they come to regret.
18:47:23	Serentty	Oh don't worry, it's not about case insensitivity. Although I do think that that opens a whole Unicode can of worms that would be better left unopened, this is about something else.
18:49:10	leorize	I think the forum is a better ground for long discussion, although it's fine here
18:50:28	Serentty	Okay, so here's my bikesheddy (but still, I believe, important) thing. The name “rune” for a Unicode code point is, I believe, even more misleading than “character”, despite the fact that it was probably chosen to avoid such connotations. The name “rune” implies something more like a grapheme cluster, and doesn't accurate describe things like whitespace, control characters, combining characters, and countless other
18:50:28	Serentty	top of that, there's also “Rune16”, which could be a surrogate half, which is pretty far from what “rune” implies. I think the current naming makes it all to easy to assume that a “rune” is something that it is not.
18:50:52	narimiran	there's already open issue about that name
18:50:55	Serentty	*accurately describe
18:51:45	lqdev[m]	is there a way of checking if a package is available during compile time? I know `pkg-config` is an option but can't find a way to check if PulseAudio is present (`pkg-config pulse` always errors)
18:51:52	FromGitter	<jrfondren> who approaches anything dealing with Unicode and says, "I bet this has no caveats." ?
18:52:01	lqdev[m]	I only need to do this on linux to get the available audio backends
18:52:06	leorize	lqdev[m]: it's libpulse
18:52:16	*	lritter quit (Remote host closed the connection)
18:52:40	leorize	the "proper" way to detect pulse is `pulseaudio --version`
18:52:50	leorize	but you'll end up using libpulse anyway, so...
18:53:13	lqdev[m]	nope, still errors
18:53:26	leorize	then maybe you don't have pulseaudio instaled?
18:53:33	lqdev[m]	aah
18:53:49	lqdev[m]	makes sense
18:54:40	lqdev[m]	I do, `pulseaudio --version` works
18:55:07	leorize	if you're on debian-based distro, you'd need libpulse-dev or similar for pkg-config to detect libpulse
18:55:23	skrylar[m]	oh boy unicode
18:55:23	lqdev[m]	how do I check JACK and ALSA? I guess `pkg-config` is my only option?
18:55:24	skrylar[m]	i sure don't have enough nightmares writing code for that \o/
18:57:30	leorize	lqdev[m]: alsa is built-in for all linux, but since you'll likely ending up using alsa-lib, check for libalsa
18:57:53	Serentty	namiran: I searched the issues on GitHub, but I couldn't find it as there are all sorts of other issues that mention runes but are unrelated.
18:59:05	narimiran	Serentty: you're right, it is not. i thought about this one: https://github.com/nim-lang/Nim/issues/7740
18:59:56	skrylar[m]	actually
19:00:14	skrylar[m]	would the old `when compiles` work with using gulp to check pkgconfig?
19:01:21	Serentty	Can I open an issue then?
19:01:32	leorize	Serentty: sure, on the nim-lang/rfcs repo please
19:01:49	leorize	skrylar[m]: you mean checking if pkgconfig exists?
19:05:53	*	Trustable joined #nim
19:12:06	lqdev[m]	how does CMake's find_package actually find a package?
19:12:40	lqdev[m]	I'm looking at this and thinking how it actually works https://github.com/andrewrk/libsoundio/blob/master/CMakeLists.txt#L47
19:13:19	leorize	those packages install .cmake files in specific locations
19:13:28	leorize	cmake have a list of them and just look up based on that
19:13:51	lqdev[m]	aah, so it's no use for me then
19:15:21	dom96	100 points, 3 hours, off the front page
19:15:25	dom96	Yep, we've been dragged down
19:15:48	dom96	Either by HN's automatic brigading algorithm or because one of the HN mods decided to do it
19:15:52	dom96	it's likely the former
19:15:58	dom96	def-, maybe you could email HN?
19:15:59	narimiran	yay
19:16:36	dom96	narimiran: why is https://nim-lang.github.io/Nim/httpclient a valid path? Why does it work without the .html?
19:17:23	leorize	dom96: github magic
19:18:33	dom96	That's bad practice, you shouldn't have the same content duplicated across two different paths like that
19:19:52	leorize	you'd have to complain with github :p
19:25:36	FromGitter	<kaushalmodi> I want to parse the timezone from a date string
19:25:43	FromGitter	<kaushalmodi> I tried: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5cf968b6702b7e5e763e7212]
19:25:59	FromGitter	<kaushalmodi> but it is printing ⏎ ⏎ ```2019-06-06T14:17:43-04:00 ⏎ LOCAL``` [https://gitter.im/nim-lang/Nim?at=5cf968c76fc5846bab733494]
19:26:08	FromGitter	<kaushalmodi> so the original timezone is lost
19:26:37	FromGitter	<kaushalmodi> if I look at the `parse` signature https://nim-lang.github.io/Nim/times.html#parse%2Cstring%2CTimeFormat%2CTimezone%2CDateTimeLocale
19:26:51	skrylar[m]	leorize: no i mean if the package can't be found in pkg-config it should fail, which means the `when compiles` fails, which should disable the block that says to set up that audio backend
19:27:05	FromGitter	<kaushalmodi> I see that the `zone` arg defaults to local() ..
19:27:15	leorize	skrylar[m]: just use staticExecEx :p
19:27:23	FromGitter	<kaushalmodi> the problem is that I need to parse the original zone from the `dtStr`.. how do I do that?
19:27:35	skrylar[m]	leorize: le shrug. it's not me who needs it :b
19:28:01	FromGitter	<kaushalmodi> @GULPF ^
19:28:11	skrylar[m]	altho thunderstrike is at the point now where i have to go dig up gtk3 and winapis to do anything more
19:30:05	Araq	Serentty, what do you suggest we use instead?
19:30:18	Araq	we need a better short, sexy term then
19:30:26	Araq	and Rune is what Plan 9 used btw ;-)
19:30:56	FromDiscord_	<DeltaPHC> `Codepoint` isn't as short or sexy, but it's quite accurate :P
19:31:20	skrylar[m]	thats what i used
19:31:46	skrylar[m]	well and a grapheme type which was a seq32 of code points (because the stream-safe encoding has a max of 30 diacritics)
19:32:01	skrylar[m]	the grapheme was mostly used in a code point interator
19:34:19	dom96	I emailed HN
19:35:19	Serentty	Araq: I was writing out a long post on GitHub. I don't have an easy answer to this. The name “code point” is the most intuitive at describing what it is to someone who already knows about Unicode, but to someone who doesn't, it's much less so. Either way, programmers have to learn something new. If “code point” is chosen, many programmers have to go look up the official Unicode terminology. If some other name is chosen,
19:35:20	Serentty	programmers have to learn the correspondences between what the programming language calls something and what Unicode calls it. I personally lean towards going with the official Unicode terminology, but I don't know of any programming language that has done that before, so it would certainly be a bold move.
19:36:43	Serentty	A much easier and more obvious decision, in my opinion, is to at least change the name of `Rune16` to `CodeUnit16` or something similar, because it absolutely does *not* represent anything similar to what `Rune` represents.
19:37:47	*	nsf quit (Quit: WeeChat 2.4)
19:37:57	dom96	Araq, You might want to reply here: https://news.ycombinator.com/item?id=20117175 :)
19:38:04	dom96	And talk about what you're working on
19:38:18	dom96	(AndyKelley is the person behind Zig)
19:38:57	FromDiscord_	<DeltaPHC> IMO, choosing the more technically correct name (Codepoint or CodeUnit) falls in line with the fact that, for example, Nim calls it "procedures" instead of going with 'functions' like any other language. Or how it's called a Seq instead of a Vector, which is technically more of a math term, etc etc
19:39:40	*	Vladar quit (Remote host closed the connection)
19:40:01	Serentty	That's actually a pretty good point I think.
19:41:26	Serentty	Okay, here's what I wrote on GitHub.
19:41:27	Serentty	https://github.com/nim-lang/RFCs/issues/151
19:49:00	*	lritter joined #nim
20:00:44	*	clyybber joined #nim
20:05:58	dom96	narimiran: some praise for your changes: https://www.reddit.com/r/programming/comments/bxhw0c/nim_version_0200_is_released/eq7i855/. Nice :D
20:06:05	dom96	I must admit, it has grown on me
20:08:29	Zevv	Same here. Although I still find bullet #1 from "Elegant" a bit patronizing and not right
20:08:57	narimiran	yaaay :)
20:09:44	dom96	Zevv, agreed
20:10:17	dom96	Please PR it if you can make it sound better: https://github.com/nim-lang/website/blob/master/jekyll/index.html#L40
20:10:37	shashlick	btw, i was unable to get jekyll serve working on ubuntu yesterday
20:10:54	narimiran	shashlick: it works for me on manjaro
20:11:02	shashlick	Key related_posts cannot be set in the drop.
20:11:06	shashlick	on ubuntu
20:11:15	dom96	did you install the Gemlock'ed deps?
20:11:28	narimiran	shashlick: `bundle exec jekyll serve`, or some other command?
20:12:55	shashlick	http://ix.io/1L6f
20:13:14	narimiran	btw, if the only criticism is just one bullet point, then we managed to improve the new website quite a lot from the initial version :))
20:13:42	Zevv	dom96: I already gave my input to Narmiran. A lot made it into the final version, but not that :)
20:14:15	*	deech joined #nim
20:16:00	shashlick	meh i give up, later
20:21:26	dom96	Zevv, nice, thank you for providing feedback. I wish I had more time to do that :)
20:23:07	FromGitter	<xmonader> @PMunch that indeed was the note i was referring to https://nim-lang.org/docs/httpclient.html#timeouts
20:24:07	Serenitor	is there anything new in regards to embedding nim script?
20:24:14	Serenitor	I see the the compiler and vm code has been cleaned up a lot. but there aren't any new helper functions to set this by any chance?
20:24:34	Serenitor	*set this up
20:25:01	FromGitter	<xmonader> also i think programming languages and compilers should have their own news websites for fruitful discussions , it's unfair lots of effort don't go to the frontage because of a new js framework
20:27:37	Araq	Serenitor, look at the tcompilerapi test for the API
20:27:40	*	narimiran_ joined #nim
20:28:43	Serenitor	ok thanks for the hint
20:29:18	Araq	also about Rune16, that's pretty easy to change we can keep the Rune16 as a deprecated alias
20:29:30	Araq	however, that is exactly what needs to stop IMO.
20:29:44	Araq	pointless renaming and redefinitions
20:30:04	*	narimiran quit (Ping timeout: 258 seconds)
20:30:42	Cadey	congrats on 0.20.0!
20:31:10	PMunch	narimiran_, https://old.reddit.com/r/programming/comments/bxhw0c/nim_version_0200_is_released/eq7i855/
20:31:19	PMunch	Reddit appears to like the new website :)
20:31:39	narimiran_	PMunch: ...and i like reddit today :)
20:31:42	Zevv	woohoo, arm builds in the nightlies \o/ when was that done?
20:32:08	FromDiscord_	<lmariscal> Araq, I agree but to be honest Rune doesn't express a lot in its own
20:32:59	Zevv	I don't really see the problem. If the documents state that "Rune" is nims name for a unicode point, then what's the deal?
20:34:10	Serentty	I mostly agree about renaming and redefinitions, but that's why I support just going with the Unicode terminology of trying to come up with some clever name that people will still disagree over.
20:34:54	FromDiscord_	<DeltaPHC> It's not a big deal in the grand scheme of things, but the rationale in this case is to not get stuck with a potentially regrettable name when we're close to 1.0 stabilization
20:35:36	FromDiscord_	<DeltaPHC> Because after 1.0, it's going to stick around for a very very long time
20:35:58	FromGitter	<GULPF> part of getting close to 1.0 is not renaming things every week imo. Nim has been "close" to 1.0 for several years now
20:36:40	lqdev[m]	I don't see anything wrong with the current name. It's shorter than "Codepoint", and for me its name resembles some "complex" character (which is exactly what UTF-8 is, encoding for characters more complex than ASCII)
20:37:12	clyybber	I agree, I like Rune, its short and expressive
20:37:15	Serentty	<Zevv "I don't really see the problem. "> Zevv: Because the usual meaning of the word is something much more like “glyph”, and having such a name gives the impression that it's actually a type for grapheme clusters. Not everyone is going to read the documentation for every type, especially if the name seems “obvious”.
20:38:20	FromDiscord_	<DeltaPHC> You could also make the argument that a name like "Rune" -isn't- obvious, and so they'll look at the docs and see that it's unambiguously a codepoint
20:39:18	rayman22201	https://memegenerator.net/img/instances/78250574/are-you-ready-to-bikeshed.jpg
20:39:35	*	rayman22201 can't help myself
20:39:54	FromDiscord_	<DeltaPHC> Personally I'm not too hung up about Rune, but I do understand the point being made
20:40:44	FromDiscord_	<lmariscal> Same here, but because Rune has already been stablished, you have to outweigh the negatives of changing it a this stage
20:40:44	Serentty	<clyybber "I agree, I like Rune, its short "> lqdev: That's exactly my point. It gives that impression, but the impression is wrong. It doesn't have to be a character at all.
20:41:09	Serentty	I meant to ping lqdev.
20:41:24	Araq	as I said, it's not even my invention, I took it from Plan 9
20:41:41	lqdev[m]	then `char` should be renamed too, if it "can be a character but doesn't have to be"
20:41:45	Araq	I think we simply need to ensure it's documented well
20:41:51	FromDiscord_	<lmariscal> ^^
20:41:56	Araq	btw we updated to Unicode 12.0
20:42:03	Araq	is that in the changelog?
20:42:06	PMunch	!eval echo NimVersion
20:42:10	NimBot	0.19.6
20:42:14	PMunch	:(
20:42:27	PMunch	It should auto-update whenever the docker images update
20:42:28	Serentty	lqdev: To be honest, I would support that as well.
20:42:33	FromDiscord_	<lmariscal> Nope, not in the changelog
20:44:00	*	narimiran_ is now known as narimiran
20:44:24	FromGitter	<GULPF> @kaushalmodi re timezone parsing: there's no way to do that in the times module
20:44:49	Araq	GUPLF: what do *you* think about breaking compiler fixes?
20:46:24	FromDiscord_	<DeltaPHC> If the biggest criticism is that Rune is named badly (as opposed to how the module implements unicode), then I'd say that things aren't too bad :P
20:46:58	Serentty	Anyway, I'm more concerned about Rune16, since it doesn't mean a code point at all, when “rune” is being established as meaning a code point. I'd be satisfied if the “rune” name were kept, but used consistently to refer to full code points and not just UTF-16 code units, which should have a much clearer name since confusing code points with code units continues to cause all sorts of bugs with non-BMP characters.
20:47:40	*	cavariux joined #nim
20:48:24	Serentty	So, forget about the rune thing. The name is fine I suppose. Rune16 is the bigger problem.
20:49:04	Araq	Rune16* = distinct int16 ## \
20:49:04	Araq	## Type that can hold a single UTF-16 encoded character.
20:49:04	Araq	##
20:49:04	Araq	## A single Rune16 may not be enough to hold an arbitrary Unicode code point.
20:49:13	narimiran	Serentty: count the amount of usages of Rune16 in unicode.nim ;) :P
20:50:36	Araq	it's not even used... wtf
20:50:48	Araq	but exported, yay
20:51:01	FromGitter	<GULPF> Araq: I think it's unrealistic to assume that everything can be fixed before v1.0, so there needs to be a way to release breaking changes in the future as well. But the promise of v1.0 stability doesn't mean much if v1.1 is released a couple of months later with a bunch of breaking changes.
20:51:27	Araq	that's not how it works
20:51:43	FromDiscord_	<lmariscal> GULPF that's why LTS versions exist
20:52:00	Araq	the point is not that we delay 1.1 or 2.0 simply to be more "stable" than we are
20:52:05	*	krux02 quit (Remote host closed the connection)
20:52:17	Araq	the point is that we backport bugfixes to 1.0.x for years to come
20:52:29	Araq	we did that with 0.19.x
20:52:40	Araq	and now the 0.19.x line is dead
20:52:53	Araq	0.20.x will live much longer
20:53:09	shashlick	@Zevv arm builds have been around for more than a month
20:53:13	Serentty	Well, “can hold a single UTF-16 encoded character” is just as inaccurate as saying that byte can hold a single UTF-8 character. And I still think it's a bad idea to use the same name for code points and code units when this exact confusion is the most common source of bugs related to surrogates.
20:53:20	FromGitter	<kaushalmodi> @GULPF Oh, I just posted https://forum.nim-lang.org/t/4912 with all details
20:53:21	shashlick	Curious on your feedback
20:53:37	Araq	also, I have a pretty good overview about what is broken in the compiler
20:53:40	FromGitter	<arnetheduck> we were just discussing if it's possible to get some more backports on 0.19.x branch while we wait for 0.20 to stabilize...
20:54:01	Araq	I went over the full list of bugs multiple times
20:54:12	FromGitter	<kaushalmodi> @GULPF I thought of using `newTimeZone` to first derive my own time zone but it doesn't work as the procs in args need to be gcsafe
20:54:17	Araq	and there are mostly 3 different pain points, all taking weeks to fix
20:54:23	Araq	or months.
20:54:59	Araq	however, it can also be documented as a list or errata like (v1 is known to depart from the spec in a,b,c,d ways)
20:55:58	FromGitter	<kaushalmodi> @GULPF So I cannot even make this work: http://ix.io/1L6u/nim
20:56:31	dom96	ooh, we can get a sponsor button on GitHub now
20:56:32	*	dom96 adds
20:57:02	Araq	arnetheduck: well ok, we can think about a 0.19.8, but it should die soon or else we rebrand 0.19 as v1
20:58:19	*	andremedeiros quit (Quit: WeeChat 2.4)
20:58:37	FromGitter	<GULPF> I think a long lived 1.0.x branch + breaking releases on v1.x/v2 like you say will be good, but I think some people are expecting a stable ecosystem and not only a stable 1.0.x.
20:58:51	FromGitter	<arnetheduck> @Araq cool - we'll ask for it only if it's important.. we have a pretty important release coming out in october that we don't want to upset too much so..
20:59:17	Araq	btw Rune16 is used exactly 0 times in Nim's stdlib
20:59:41	Araq	I think we can savely deprecate this for v1, rules exist to be broken
21:00:51	Araq	GULPF: for a stable ecosystem packages should support v1.0 and not "devel only"
21:01:13	FromGitter	<alehander42> nice
21:01:16	FromGitter	<alehander42> congrats
21:01:17	FromGitter	<alehander42> 1) 20!
21:02:04	Serentty	Okay, so then instead of renaming it, it could just be removed, with int16 becoming the recommended type to hold a single UTF-16 code unit? I would be fine with that.
21:02:18	FromGitter	<arnetheduck> one fairly core thing to consider for deprecations could be to come to a conclusion about type-based signed/unsigned vs operation-based.. ie `+%` and friends
21:03:02	Araq	the conclusion was made, 'shr' changed to be type-based
21:03:10	FromGitter	<alehander42> btw why
21:03:17	FromGitter	<alehander42> no hot code reloading and new runtime
21:03:19	FromGitter	<alehander42> in the blog
21:04:00	*	narimiran quit (Ping timeout: 248 seconds)
21:04:02	FromGitter	<arnetheduck> `int16` seems like a tricky type to use for utf-16, given that it's a bit-based thing / kind-of-unsigned..
21:04:18	Araq	well it should be uint16, yeah
21:04:35	FromGitter	<arnetheduck> so maybe get rid of op-based then, just creates two ways to do the same..
21:04:57	FromGitter	<GULPF> @kaushalmodi: newTimezone accepts closures, so you can make your `newTimezone` example work that way. My `timezones` package have a proc for doing the offset string -> Timezone conversion I think
21:05:13	Serentty	I personally prefer separate types for signedness, because otherwise it's really easy to misinterpret one as the other.
21:06:11	Araq	I'm still a bit torn but the fact is that not even Araq-himself could implement MD5 with signed ints and unsigned operators
21:06:30	Araq	and so it's a failed experiment ;-)
21:06:30	FromGitter	<GULPF> @kaushalmodi https://github.com/GULPF/timezones/blob/0c5f270c81b798d16b4dc690bbfc0c22fcfa37ea/timezones.nim#L101
21:06:42	Serentty	For UTF-32, signedness intentionally doesn't matter, but it does for UTF-16, yeah.
21:06:52	FromGitter	<arnetheduck> it's kind of useful that `%` works as expected for unsigned ints
21:06:53	*	Trustable quit (Remote host closed the connection)
21:07:04	FromGitter	<kaushalmodi> @GULPF Thanks! I will install `timezones` then
21:07:33	FromGitter	<arnetheduck> there are some benefits to op-based - it's easier to create a compiler for example because computers are op-based
21:08:18	Araq	it's good for an IR
21:08:25	Araq	but pretty bad for humans
21:08:44	Serentty	Araq: Ah, so has your opinion on unsigned types changed since the posts from a few years ago that I read?
21:09:14	FromGitter	<arnetheduck> yeah, that sums it up pretty neatly
21:09:15	*	envoyt quit (Ping timeout: 248 seconds)
21:09:53	Araq	not really, Serentty. For example 'len' should really return a signed number because then stuff like len - 4 > y works
21:10:11	dom96	There we go: https://nim-lang.org/blog/2019/06/06/version-0200-released.html#installing-0200
21:10:17	*	envoyt joined #nim
21:10:19	dom96	Hope that makes sense
21:10:21	Araq	however, unsigned is acceptable if you remove the minus operator from it
21:10:54	Araq	unsigned simply has no minus operation, too error prone, you always need to use i+4 < len instead of i < len-4
21:11:21	Araq	but I have yet to see anybody propose this solution :P
21:12:42	FromGitter	<arnetheduck> we've proposed it, for the ethereum 2 specification :) it's happening there
21:13:03	Araq	wow
21:14:13	FromGitter	<kaushalmodi> @GULPF got it working.. thanks!!
21:14:17	FromGitter	<kaushalmodi> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5cf98229faac64393450567f]
21:15:15	dom96	arnetheduck: Araq: nice, I like that solution
21:15:34	dom96	We gonna get Ethereum contracts written in Nim?
21:15:42	FromGitter	<arnetheduck> it's been a long road there however, most of the spec people are python devs armed with `BigInt`.. even sha256 hashes are signed numbers in their world
21:16:41	FromGitter	<arnetheduck> dom96, possibly - we're having a community call about it on friday
21:16:43	skrylar[m]	maybe etherium 2 won't have the bug that the developers tampered with history for personal gain thus ruining the whole thesis of etherium in the first place :^)
21:16:53	dom96	arnetheduck: anything we can do to help?
21:17:39	Araq	alehander42: I think we should write separate articles about the new runtime and hot code reloading
21:17:58	Araq	and before that I need to try hot code reloading too, never used it
21:18:15	Araq	don't have time to develop games in Nim...
21:18:28	dom96	I'm developing a game in Nim :P
21:18:37	FromGitter	<arnetheduck> but this is outside of contracts - it's about verifiability in the spec for the state machine that governs consensus.. life is easier if you can take underflows out of the equation when using unsigned arithmetic.. where possible we're going for saturation as well
21:18:50	dom96	But it runs in the browser so not sure how much this new runtime would help me
21:19:00	dom96	what would help me is reducing the overhead of all these string conversions :/
21:19:16	dom96	I might have to end up writing JS string utils
21:19:16	clyybber	webAssembly would help you :P
21:19:19	Araq	dom96, I simply use a different coding style in karax for that
21:19:27	Araq	it works
21:19:27	FromGitter	<xmonader> would hot code reloading mean I can register/unregister plugins in the runtime without having to recompile the app?
21:19:35	clyybber	yes
21:19:41	clyybber	well, you recompile it
21:19:48	Araq	clyybber! where is my =move?
21:19:48	dom96	Araq, the whole idea is that I want to reuse code on the front end and the back end (it's a multiplayer game with both client and server simulating the world)
21:20:09	Araq	dom96, kstring is a slim abstraction for that
21:20:11	clyybber	Araq: just started working on it again :#
21:20:28	Araq	come on, it's easy, writing the spec was the hard part
21:20:36	Araq	;-)
21:20:40	clyybber	yeah, :)
21:20:42	FromGitter	<arnetheduck> I've been thinking that it would make some sense to refactor the primitives in system.nim a bit and put all of them in their own file, with functions covering overflow-exception, saturation and wrapping variations, then mapping `+` to the right one depending on compiler options and that stuff.. would be nice if the `--checks:on/off` could be moved to the std lib a bit more instead of sitting in the compiler
21:20:47	skrylar[m]	how does a 5..0 counter work with no minus operator :f
21:21:20	clyybber	with signed integers?
21:22:04	*	couven92 joined #nim
21:22:07	FromGitter	<xmonader> @Clyybber ybber how is that `hot reload` if i need to recompile it?
21:22:13	dom96	Araq, Maybe that should be its own Nimble package? :)
21:22:18	FromDiscord_	<DeltaPHC> Vladar: Looks like you might wanna push out a bugfix for sdl2_nim. The integer conversion error means it doesn't compile by default on the Nim 0.20.0 release
21:22:24	FromGitter	<jrfondren> you still need to compile the thing you're loading.
21:22:36	FromGitter	<arnetheduck> @dom96 well, one of the things we have to consider is whether to write a new `system.nim` completely for contracts/wasm or go with the present one, which has a bit too much stuff for that environment
21:22:52	skrylar[m]	i suppose the nice thing is with distinct types you can make super paranoid counters that do all of this (or atomics too, fun times.)
21:22:56	FromGitter	<xmonader> yeah I'll compile all of my plugins into .so and then can load/unload them in the app entry point
21:22:58	Araq	write a new system.nim
21:23:04	FromGitter	<xmonader> e.g webserver adding modules to it
21:23:23	clyybber	xmonader: It reloads modules while the program is running?
21:23:42	dom96	arnetheduck: do you have an idea of what you want out of system?
21:23:48	clyybber	Thats the definition of HCR in my book
21:23:48	FromGitter	<xmonader> or the idea i had long time ago, something like ansible in nim, and separate compiles of plugins without actually having to update the binary itself
21:24:18	FromGitter	<arnetheduck> it's easier to enumerate what I want *in* system :)
21:24:31	FromGitter	<xmonader> @clybber which book is that?
21:24:35	dom96	then yeah, you probably want your own system :
21:24:36	dom96	:)
21:24:56	*	alexander92 joined #nim
21:26:20	Araq	yup, write your own system.nim. Best reason to do it: It's never been done before.
21:26:57	Araq	if that's not fun then I don't know what is. :-)
21:27:11	alexander92	i planned to do something like that(maybe) for my kernel
21:27:25	alexander92	but i decided to just use c until i actually get an idea what i need
21:27:44	clyybber	xmonader: my book :p
21:28:15	clyybber	as in my opinion
21:28:36	FromGitter	<xmonader> Aha, not native speaker sometimes I take things literally. Sorry.
21:28:43	PMunch	Can a macro have a default parameter?
21:28:47	clyybber	np :), not a native speaker either
21:29:20	PMunch	Or rather, how do I check if the parameter is set or not
21:29:53	alexander92	i think i used nil sometimes PMunch
21:30:09	alexander92	but not sure if thats smart
21:30:17	PMunch	Hmm, it just tells me it's not able to infer the type..
21:31:21	PMunch	Ay, I think I got it working :)
21:31:38	PMunch	Just as I realized that I don't need this..
21:32:46	*	jasper_ quit (Quit: Page closed)
21:40:28	deech	Araq, just saw the release announcement and didn't know it was a 1.0 candidate. I'm working on https://github.com/nim-lang/Nim/pull/11341 which I think should be in place for 1.0. It basically fixes a lot of the compile/runtime interaction issues with assignments and operands. It's not quite ready yet, what is your 1.0 timeline?
21:40:53	Araq	I won't tell because you'll laugh
21:40:59	deech	Ok good.
21:41:17	Araq	but I know you're working on this and in my mind it ships with v1
21:41:33	deech	Awesome, appreciate it.
21:41:57	Araq	however, please also update the spec while you're at it
21:42:14	deech	Ah, ok. I forgot about that. I did update the manual.
21:42:22	Araq	the manual is the spec.
21:42:26	Araq	:-)
21:44:15	*	a_b_m joined #nim
21:45:11	*	envoyt quit (Ping timeout: 252 seconds)
21:45:59	*	alexander92 is now known as rename
21:46:13	*	rename is now known as alexander92
21:46:46	alexander92	hm Araq why is your name with `@` in irc
21:46:54	alexander92	ah this is some admin stuff
21:46:59	alexander92	right
21:47:39	*	abm quit (Ping timeout: 248 seconds)
21:48:01	PMunch	Hmm, this RPN calculator thing I've been toying with is actually becoming pretty neat
21:49:00	alexander92	this is how your new language starts
21:49:22	PMunch	Oh, it's already started :P
21:49:24	alexander92	you need just a http and GET operators
21:49:38	alexander92	and it will suddenly become a backend hotness
21:49:52	PMunch	I have commands, which play back what you store in them. And it has goto, loop counters, and logic operators :)
21:50:14	alexander92	so you're turing complete already
21:50:21	PMunch	The whole thing is built with only two data-types though, float and "label" which is just a string without spaces
21:50:40	alexander92	i think you need a term rewriting mechanism
21:50:48	PMunch	Term rewriting?
21:50:49	alexander92	and you'll be able to metaprogram your own calculator
21:51:47	alexander92	https://homepages.cwi.nl/~daybuild/daily-books/learning-about/term-rewriting/term-rewriting.html
21:52:36	alexander92	afaik just a way to define rules how an expression is rewritten into a new expression
21:52:45	*	Summertime quit (Quit: Sunsetting.)
21:52:50	PMunch	Well you can write a command that creates a new command already
21:52:54	PMunch	I think..
21:52:58	PMunch	I haven't actually tried that :P
21:53:33	alexander92	so you can a 2 + sum2(1) define
21:53:45	alexander92	2 sum2 #
21:53:51	alexander92	yeah i remember you had something like that
21:54:15	alexander92	but i am thinking more of symbolical rewriting
21:54:33	alexander92	where your calculator can somehow work with symbols and e.g. transform relations
21:54:35	PMunch	Hmm, that didn't quite work..
21:54:40	alexander92	(a+b)^2 == (a+b)*(a+b)
21:54:44	alexander92	stuff like this
21:54:51	alexander92	but this is extremely unneeded :P
21:55:19	*	Summertime joined #nim
21:55:23	PMunch	Aaah
21:56:50	PMunch	Hmm, it didn't like making commands within commands
21:56:56	PMunch	I'll have to look into that tomorrow
21:56:59	dom96	PMunch, sounds a lot like my stack-based lang: https://github.com/dom96/nael :)
21:58:18	PMunch	Well look at you mr. fancy-pants with your many types
21:58:35	alexander92	full blown language
21:59:02	alexander92	if you dont have generic types in your calculator
21:59:09	PMunch	Yeah, I basically wanted to just make a more powerful calculator, without any mental overhead
21:59:12	alexander92	are you really doing stuff right
21:59:29	deech	Araq,currently the type checker allows a compile time type like, 'let x {.compileTime.} = "abc"' to be passed to a proc like 'proc f (a:string) ...', I would like this to only be allowed if either 'f' is '{.compileTime.}' or 'a' is a
21:59:31	alexander92	(but very nice seriously)
21:59:51	deech	'static[string]', where should I go looking for that? `proc matches` in sigmatch.nim?
22:00:06	PMunch	Hmm, can I detect if the user is typing into a terminal or if this is input being piped to it in Nim?
22:00:26	PMunch	Well, since you can in C, obviously you can in Nim
22:01:50	alexander92	hm cant bash build_all.sh
22:01:51	PMunch	Aha, terminal.isAtty
22:02:56	disruptek	i really don't understand these new object variants.
22:03:20	disruptek	why do i need to provide compile-time values for the case branches?
22:06:59	skrylar[m]	to avoid errors at runtime i guess
22:07:57	disruptek	uhhh
22:10:14	PMunch	http://ix.io/1L6T
22:10:15	PMunch	Hmm
22:10:22	sealmove	This is briefly explained in the Manual. It's lets the compiler make better guarantees and optimizations.
22:11:08	disruptek	the variants section doesn't appear to have changed much, if at all. where is this covered in the manual?
22:11:55	PMunch	Hmm, that might break nimlsp as well..
22:12:33	sealmove	"Because case statements are checked for exhaustiveness during semantic analysis, the value in every of branch must be a constant expression. This restriction also allows the compiler to generate more performant code."
22:12:42	sealmove	https://nim-lang.org/docs/manual.html#statements-and-expressions-case-statement
22:13:18	sealmove	also I just found a relevant bug
22:13:29	disruptek	this is a mess.
22:13:46	sealmove	if you use a converted for these constant expressions, and a range, then it crashes
22:14:01	sealmove	https://github.com/nim-lang/Nim/issues/11422
22:14:05	sealmove	converter*
22:14:48	PMunch	Hmm, it works somewhere..
22:14:49	sealmove	I guess because it has to convert all the intermediate values some how, and this senario is not covered
22:15:15	disruptek	pretty sure this is broken; i've got a discriminant that contributes nothing to the object, but somehow it's demanding that i add a field that, again, is not defined for this variant.
22:17:38	PMunch	http://ix.io/1L6W
22:17:42	PMunch	That doesn't work
22:19:38	disruptek	wow, you need to create the objects with constants, too. madness.
22:20:21	sealmove	lol that's obvious. you either havs CT values or not.
22:21:24	disruptek	so i parse a discriminator, then have a tree/case where i compare that value to a constant which, duh, i use to construct the object.
22:21:37	disruptek	but despite the case statement asserting equality, i cannot use the variable itself.
22:22:26	disruptek	"Often an object hierarchy is overkill in certain situations where simple variant types are needed."
22:22:36	PMunch	Nice, now my calculator can run statements I pipe into it as well :)
22:22:37	disruptek	often, it seems, you will hate yourself for choosing variant types.
22:23:08	sealmove	it takes some practice to know when you can/use use them over objects
22:23:18	FromGitter	<jrfondren> https://github.com/nim-lang/Nim/issues/11143 <- Nim doesn't use case control-flow for inference on variants
22:23:21	sealmove	can/should *
22:24:01	*	sealmove quit (Quit: WeeChat 2.4)
22:24:12	disruptek	jrfondren: i thought that's what the rewrite was supposed to fix. i'm the dummy for not following it more closely.
22:24:37	FromGitter	<jrfondren> aye I didn't notice it was supposed to be fixed
22:25:28	*	alexander92 quit (Ping timeout: 245 seconds)
22:26:13	disruptek	fool me once, shame on you, fool me twice, shame on me.
22:26:42	*	PMunch quit (Remote host closed the connection)
22:27:39	*	couven92 quit (Quit: Client Disconnecting)
22:32:07	*	deech quit (Ping timeout: 252 seconds)
22:36:04	Araq	why is #11143 still open?
22:36:09	Araq	I thought it was fixed...
22:41:28	*	deech joined #nim
22:46:38	*	I_Right_I joined #nim
22:56:15	*	clyybber quit (Quit: WeeChat 2.4)
22:58:23	*	lritter quit (Quit: Leaving)
23:01:22	*	luis_ joined #nim
23:06:03	*	a_b_m quit (Ping timeout: 248 seconds)
23:08:56	*	laaron quit (Quit: ZNC 1.7.1 - https://znc.in)
23:08:57	*	laaron- joined #nim
23:10:35	*	laaron- quit (Remote host closed the connection)
23:10:56	*	laaron joined #nim
23:17:52	*	luis_ quit (Remote host closed the connection)
23:26:43	*	a_b_m joined #nim
23:34:44	*	I_Right_I quit (Remote host closed the connection)
23:57:17	*	a_b_m quit (Quit: Leaving)